pchalasani/claude-code-tools

Add Pattern records newly learned Langroid design patterns into the patterns library with a reference link and an implementation document. Use it when you discover a pattern to keep the catalog organized and searchable for future design problems.

Agent Style applies 21 literature-backed English technical-prose writing rules (passive voice, concrete language, needless words, citation discipline) for AI-assisted formal documents. Use when the user explicitly asks for agent-style by name for papers, design docs, proposals, READMEs, or commit messages.

Avoid AI Writing audits and rewrites content to remove AI writing patterns (AI-isms) such as em-dash overuse, bold overuse, and excessive bullet lists. Use it when asked to remove AI-isms, clean up AI writing, or make text sound less like AI.

Code Walk Thru opens files in the user's editor one by one at specific line numbers so they can follow along while the agent explains code or shows changes.

Dynamic Workflow creates, runs, and manages durable JavaScript workflows that coordinate multiple headless Codex agents for fan-out/fan-in, multi-stage pipelines, and background runs. Use it when you need dynamic branching, looping, or long-running multi-agent work that a single model turn cannot handle directly.

Issue Reply Wakeup registers an existing GitHub issue so the current Codex or Claude Code session is woken when its first future comment arrives. Use when work is blocked on a GitHub reply and the session must resume asynchronously.

Langroid Patterns catalogs design patterns for the Langroid multi-agent LLM framework, including task returns, tool handlers, task control, MCP integration, and quiet mode. Use it to find reference code examples when building Langroid agents.

Log Work records the work segment since the last entry into a dated WORKLOG markdown file. Use it to keep a concise timestamped history of files created, read, or changed during a session.

Make Issue Spec creates a markdown document that clearly specifies a task, asking clarifying questions for underspecified parts and staging the file in Git when possible. Use it to define work for agents before implementation.

Msg sends messages between agent sessions running in tmux panes. Use it to coordinate with other Claude Code or Codex CLI agents by registering, sending, replying, and checking an inbox.

Recover Context pulls the last task, current state, pending items, and key decisions from the most recent parent session in a session lineage, delegating to sub-agents or a session-search skill to avoid loading large histories directly. Use it to resume prior work after a reset.

Remove AI Patterns removes AI-writing patterns ("AI-isms") from text using a vendored catalog and detector, with detect-only, edit-in-place, voice profiles, and convergence passes. Use it to audit and revise prose so it no longer reads as AI-generated.

Session Search searches previous code agent sessions for specific work, decisions, or code patterns. Use it when you need to locate past work or decisions from Claude Code or Codex CLI sessions.

Socratic Quiz guides users to deep understanding through graduated, adaptive questioning rather than direct explanation. Use it when a user asks to be quizzed, taught, or tested on a concept.

Tmux CLI lets a Claude Code agent send commands and messages to other CLI agents or scripts running in separate tmux panes. Use it when you need to coordinate work across multiple panes, run shell commands with exit code detection, or communicate with another Claude Code instance.

Voice Update uses a pocket-tts script to speak short, conversational summaries as spoken audio feedback. Use it when the agent finishes a task, is reminded by a Stop hook, or the user asks for a spoken update.

Voxtype Install guides the user through installing, configuring, and launching voxtype, a fully on-device macOS voice dictation tool that types transcribed speech into the focused app.